cv
Basics
| Name | Najwa Laabid |
| Label | ML Engineer & PhD Researcher |
| najwa.laabid@aalto.fi | |
| Url | https://najwalaabid.github.io |
| Summary | ML engineer and PhD researcher with 6+ years building end-to-end solutions across drug discovery, genomics, network security, and enterprise software. Experienced across the full stack from prototyping to deployment, with a consistent record of shipping pragmatic systems in both industry (Google, Ericsson) and research settings. |
Work
-
2022.05 - Present Doctoral Researcher
Aalto University
PhD research in machine learning applied to drug discovery, focusing on graph diffusion models for retrosynthesis.
- Published at ICLR 2025 and ICML 2024 workshops
-
2021.09 - 2022.04 Junior Researcher
University of Eastern Finland
Built a VAE model for a NeurIPS 2022 competition tackling multimodal single-cell integration, a core open problem in computational biology with direct implications for cancer disease profiling.
- PyTorch
- AWS (EC2/S3)
-
2021.06 - 2021.08 ASCI Programme - Intern
Aalto University
Implemented approximate inference for nonstationary Gaussian processes via Kalman smoothing in JAX, enabling GP models to adapt to signals with time-varying dynamics such as gravitational wave chirps.
- PyTorch
- JAX
- AWS (EC2/S3)
-
2020.06 - 2021.05 Thesis Placement - Data Analytics
Ericsson Inc.
Studied IoT botnet C&C communication across 14 malware families; built a Random Forest detection system across three traffic types, achieving near-perfect F1 on decentralized botnets and 0.85 F1 on real centralized captures.
- Python3
- AWS (EC2/S3)
-
2019.07 - 2019.09 Software Engineering Intern
Google Inc.
Built metadata parsing logic for Google's internal AV-preset automation, reducing manual meeting room setup for one of the world's largest internal conferencing tools.
- JavaScript
- HTML
- CSS
-
2019.01 - 2019.05 Project Coordinator and Developer
CLE at Al Akhawayn University
Built a web platform replacing a manual booking system, handling session bookings and hour logging across 79+ courses and 2050 students; later led the redesign.
- React
- Node.js
- HTML5
- CSS
-
2018.07 - 2018.12 Software Engineering Intern
Sankhya Jiva Tecnologia e Inovação LTDA
Replaced Dialogflow with a custom open-source temporal expression parser for Brazilian Portuguese in enterprise chatbot BIA to improve language accuracy and eliminate vendor lock-in; adopted and extended by the team post-internship.
- Java
- JUnit
- Kotlin
-
2018.01 - 2018.06 Full-Stack Developer
ASC Brazil
Developed a social media management module for an ERP platform, enabling clients to read and post across YouTube, Facebook, Instagram, and Twitter directly from within the system.
- PHP/MySQL
- NodeJS/MongoDB
- HTML/CSS
Education
-
2022.05 - Present Espoo, Finland
-
2019.09 - 2021.07 Joensuu, Finland
M.Sc.
University of Eastern Finland (UEF)
Information Technology (focus on data science, minor in statistics)
-
2015.09 - 2019.06 Ifrane, Morocco
Awards
- 2022
Finland Fellowship
7.5 month salary and relocation bonus.
- 2019
UEF IMPIT Scholarship
100% tuition and living expenses stipend.
- 2018
Women Techmakers Scholar
Google Inc. and Women Techmakers
- 2015
Al Akhawayn University Excellence Scholarship
100% tuition and fees.
Publications
-
2025 Equivariant Denoisers Cannot Copy Graphs: Align Your Graph Diffusion Models
ICLR 2025
We demonstrate that standard permutation equivariant denoisers face fundamental limitations in graph-to-graph translation tasks and propose aligning input and target graphs to break input symmetries while preserving permutation equivariance.
-
2024 Aligned Diffusion Models for Retrosynthesis
ICML 2024 Workshop on Geometry-grounded Representation Learning and Generative Modeling
-
2024 Conditional Flow Matching for Time Series Modelling
ICML 2024 Workshop on Structured Probabilistic Inference & Generative Modeling
Skills
| Machine Learning | |
| PyTorch | |
| JAX | |
| TensorFlow | |
| Scikit-Learn |
| Software Engineering | |
| Python | |
| Java | |
| JavaScript | |
| React | |
| Node.js | |
| Flask | |
| Django |
| Infrastructure | |
| AWS (EC2/S3) | |
| OpenShift | |
| Docker |
Languages
| Arabic | |
| Native speaker |
| French | |
| Professional proficiency |
| English | |
| Professional proficiency |
| Portuguese | |
| Intermediate |
| Finnish | |
| Beginner |
Interests
| Machine Learning | |
| Graph Diffusion Models | |
| Drug Discovery | |
| Retrosynthesis | |
| Generative Models |
| Applications | |
| Computational Chemistry | |
| Genomics | |
| Single-cell Analysis |
Projects
-
RxnLab: Open platform for retrosynthetic analysis
Open platform bridging ML research and chemistry practice, hosting retrosynthetic models on CSC's Rahti infrastructure.
- Flask
- RDKit
- PyTorch
- JavaScript
- OpenShift
-
Denoising scRNA-seq data using Bayesian Inference
Imputed dropout events in scRNA-seq data by fitting a Zero-Inflated Negative Binomial noise model via deep autoencoders.
- Python3
- TensorFlow
- Scikit-Learn
-
Adversarial Resistance in Toxic Comments Detection
Showed that small-scale adversarial training improves robustness of toxic comment classifiers without degrading accuracy on clean data.
- Python3
- PyTorch
- Scikit-Learn
-
Staffing Plan App for Jacobs Engineering (JESA) Group Inc.
Web-app managing resource allocation and project development for JESA Morocco.
- Django
- JavaScript/jQuery
- HTML5/CSS